"As a statistician, I have a problem with the sigma metric..." onesixsigma.com interviews Jeroen de Mast

onesixsigma: When did your interest in Six Sigma start?

Jeroen de Mast: I was working for my current employer and at the time our main business was statistical process control, back in 1995. Then, General Electric Plastics, who have their European headquarters in the Netherlands, wanted to start a Six Sigma initiative. They were looking for a consultant who could do the training for them. That’s how we got involved, seeing the training material that GE Plastics was using, we were actually very enthusiastic about it.

After GE, DAF trucks, a Dutch truck manufacturer but part of PACCAR, an American manufacturer, wanted to start Six Sigma; that was our second customer.

And then the next step was for us to introduce Six Sigma at Douwe Egberts, the coffee company. We mainly got involved via Dutch divisions of American companies that were starting Six Sigma.

You have written three books on the subject of Six Sigma. What inspired you to initiate the research which led to these books? What was the trigger? Was it a lack of books?

In the beginning it was a lack of books, we wrote the first book in 1999. At that time there were actually only a few books, and they were not suited as a basis for training. It is practical to have a book if you deliver training for instance. Most of the participants like to have a book which they can show to their boss, or can read at their own speed at home.

That is actually how we started the first book. And the second one is the Dutch “Stap voor Stap” book, the 3rd one is on Lean Six Sigma for Service and Healthcare. Of course there are a lot of good books now, and I don’t particularly like writing books! I prefer writing short articles.

But on the other hand, we are working for the university and I have a position at the university as associate professor. We have a strong ambition in doing academic research on this topic. I think we have some nice things to say on what we have been working on for the past few years and that was actually the motivation to write the books. We aim for the books to be quite short and to the point.
It is not really an academic book, but it is not a flimsy book either; that was the intention.

And it does go further; it is not just an introduction.

Our background is Mathematics and Statistics. We also developed an interest in Business Economics and even Philosophy of Science. That all comes together in Six Sigma. And we were aiming to write a book that is - if you look at the business economic side of Six Sigma - not just all wild claims, but actually relates to the paradigms which you can find in the business economic .

That was the aim of the book: write a book grounded in sound academic research and to the point.

It is unusual that a university who is supposed to be unbiased works with commercial companies. You do want to keep your scientific, academic view?

There is a strain there: we are a commercial firm, we make profit. We make a separation; we have our academic research papers, which are peer reviewed, objective. And on the other hand we have a commercial business: which provides training, consulting. The way we present Six Sigma when we try to find new clients, is different of from what we say in academic papers.

But the difference is not that big. Simply, we believe in Six Sigma. But I don’t believe in everything of Six Sigma. I find the business economic underpinning given in many accounts quite weak. That is not the story we give: what we believe is sounder economic underpinning.

For instance as a statistician, I have some problems with the sigma metric: I find the metric simply a very difficult and complicated way to express a percentage of non-conformation, you can hardly make any mistakes estimating that one.
If you try to express that number in the form of the Six Sigma metric, that works in the case of a 100% standard situation where you have a normal distribution, and at least one specification limit.

But if you are in a non-standard situation with non-normal data or you don’t have a sensible specification limit, there is a large number of possible modifications, and there are many mistakes you can make to calculate the Six Sigma level. The problem is, you have to make choices then and in the end it makes it rather arbitrary how you calculate it, and based on the definition you use, you get quite different results.

So you don’t use it?

No, we explain it to our customers, they should be aware of it and how you calculate it. It is a coincidence that the name is based on this aspect. But I don’t think it is an important aspect of Six Sigma. I find the tools, the DMAIC structure and the organisational structure way more important. All of those make sense.

There is a strain between the commercial point and the academic point of view. Of course the language is a bit different, and the emphasis is different. If you compare the organisational structure that Six Sigma proposes to recommendations made in economic organisational theory, that matches. As an academic I can defend that and I can explain why. Of course the names ‘Black Belt’ and ‘champion’ are not used in the economic literature, but if you look at the principles behind it, you see a match there. And that is why, as an academic, I can sell it if I work as a commercial person.

What do you think about certification?

When we started in the Netherlands we found those names Black Belt and Green Belt a bit funny, our assessment was: it is too American, we thought Dutch companies would prefer more modest terminologies, like ‘Project Leader’. But we were wrong. It really works, simply because the names ‘Black Belt’ and ‘Green Belt’ are recognised globally. Each company has its own criteria, though, as to what counts as a Green or Black Belt.

The University of Amsterdam has an official university diploma, so there is a examination committee. You have to do an exam covering the theoretical part and you have to submit 2 projects. Most serious companies have similar criteria.

Also if you look at the training curricula, among the serious literature you see a lot of consensus covering the criteria. However, a lot of consultancies that do not really want to step into Six Sigma, are selling a lot of products like Lean for which they use Six Sigma terminologies. For instance you can buy Six Sigma Black Belt training covering three days and you get a Black Belt certification. But I don’t think that is Six Sigma. That is simply consultants using the terms Lean and Black Belt and selling it.

Just if you look at the serious providers of Six Sigma Black Belt training then you see that a typical training curriculum is between 14 and 20 days. And there is quite some consensus and understanding in what should be part of the training, and what the criteria is for Black Belt certification.

This does not hold for Design for Six Sigma by the way: there is DMADV, IDOV, IDDOV, there is a lot of variation there. In Six Sigma proper you see every body uses DMAIC. There is some difference between Six Sigma for Service and Industry. Within those fields, there is a lot of consensus which tools and approach is used in training.

Which tools and terms in specific areas are used?

If I look at our industrial clients, then Design of Experiments is by far the tool which gives you the most benefits. I really like the step-by-step strategy of DMAIC, but if you are talking about tools in industry, Design of Experiments is generally an important tool. Because in industry people know basic statistics, they know control charts and process capability analysis. But Design of Experiments, that is an eye opener for them.

For our clients in Banking and Healthcare, process capability is quite an effective tool. Our emphasis is less on statistical tools; we do a lot on effective brainstorming, autopsy, ‘BOB’ (best of the best) versus ‘WOW’ (worst of the worst). It is not rocket science. An autopsy simply means you isolate some really bad examples. For instance, if you are running a project on lead time of some back offices, then autopsies mean that you single out some of the cases that really had a long lead time and you do a close investigation of what precisely happened there.

Simple tools like that really have added value in Services and Healthcare because quantitative data has a lot of impact there. If you do a Process Capability Analysis, that really gives you the picture of where you stand. But if you want to improve and want to discover the causes, the answer is not in the data. In order to get to the answer, you have to get another sort of information, namely: bad examples and people who work in the process. They have a lot of information.

Healthcare and Finance: is that a reasonably new field for you?

No, we started in that field in 2001, 5 years ago. And the majority of our clients are now from Finance, Service and Healthcare. What we see happening in the Netherlands is manufacturing is leaving. What is remaining here is Design for Six Sigma in industry. What is staying here is product and process development. That is mainly what we do in industry now: that is Design for Six Sigma.

Do you see that evolving, in Healthcare and Services? How do you see that in the future?

Yes. If you look at how government institutions, banks, insurance companies, and hospitals have organized their processes, they are 25 years behind industry.
What you see now is they are actually running in, by copying proven principles from industry. Six Sigma entails a lot of proven principles from the 20th century. Finance, insurance companies and Healthcare are using them now to close the gap and enter the 21st century when it comes to efficiency and effectiveness of their processes. The big one missing so far is the government. But government will follow.

What do you say to critics?

Of course some criticisms are valid. There is a lot of hype, a lot of nonsense, in books. However you should read through the way Six Sigma is presented. And if you look at the methodology from the viewpoint of philosophy of science, which is one of my expertises, you see that the Six Sigma approach really matches with good science. Good empirical science follows quite similar steps as the DMAIC steps, for instance.

If you don’t mind too much the slick presentation, and you look underneath, the underpinnings, you’ll find it is grounded in good science. I think Six Sigma has borrowed a lot of proven principles from the 20th century. And these principles have somewhere along the road been grounded in science.

So my answer to critics is yes, there is a lot of hyperbole on how it is presented, but underneath it really makes sense.

What are the key deployment problems in service sectors?

Especially in Finance and Insurance there are some generic sort of projects which you encounter over and over again. The big majority of projects are about throughput time and about processing time. Throughput time is an aspect of service quality time, and processing time that relates to the capacity, to the staff they need (and thus relates to costs).

In a typical services application, the 3 CTQs will be: throughput time (which is caused mainly by waiting time), processing time or cycle time, and the amount of rework.
That is the typical project for a bank or insurance company and even in healthcare. And the solution will involve a lot of Lean principles: line balancing is particularly useful. But it is really important that the Lean tool set is placed within the DMAIC structure. Because the application of Lean tools risks being undirected and unfocused. The CTQ Flowdown and process capability analysis in the Define and Measure phase bring focus to the really important issues, while the Control phase ensures that the improvements are integrated in line management and the process.

For instance, one of the most effective tools at the back office is line balancing; where you balance the staff for each step in the process. In order to do so, you really need to know the cycle times, you have to measure them. And that is so effective because that has rarely been done before in a back office. Once you start measuring cycle times, you’ll typically find that there are process steps with substantial overcapacity. Here, Six Sigma and Lean can easily lead to important savings.

If you measure cycle times, if you study the routing in processes, if you consider the applied queue management, you see that with almost trivial instruments you can at the same time improve the through-put time by factors of 10. We’ve seen throughput times reduced from 30 days on average down to 3 days, simply by applying proper queue management and by applying a proper routing system.

The next step is measuring the cycle times and doing the calculations. An over-capacity of 30 to 50 % is more standard than the exception. So in one project you get: a dramatic through-put time reduction and, if you want that, a dramatic reduction of staffing.

That is the typical project in the service sector: in banks and insurance companies. But projects can be quite diverse. In hospitals we’ve had some projects about length of stay in a hospital, and for instance about the laundry process.

Usually the main thing in that sort of organisation is that there is not a tradition of process engineering. You have a lot of people with a commercial background or background in law or a medical background. But there aren’t that many people who are experts in how to design a process or how to manage a process, contrary to industry where you have Process Engineering. Each factory has Process Engineers.

That is where Six Sigma really helps in banks, insurance companies, hospitals: it trains Black Belts into Process Engineering: how to measure a process, how to manage a process or how to design a process.

Have you seen many instances of Lean Six Sigma applied to sales and marketing?

Not too many. The reason for that is, from my experience, those are the departments in an organization that are typically the least open for Six Sigma. If you talk to sales persons or marketing persons they are not used to see what they do as a process.

It would be interesting to change marketing processes into Six Sigma processes...

We did that for a large bank in the Netherlands. We worked quite a lot with marketing and sales there and we were quite effective there. But when starting Six Sigma in a company, I don’t start there, back office processes are easier and the people are more open there. It helps to get some successes there and then go to marketing and sales.

The problem is that Six Sigma improves everything that is done in a routine manner. And that is the definition of a process: everything you do in a routine, that is: repeated, manner. People from sales and marketing have to change their view on sales and marketing. They have to become aware that 90% of what they do is routine processes. Of course there is also some intuition and creativity involved, but 90 % is routine. And if it is a routine you can collect data from it and then you can improve. But that is a prospect that they resist. It is the same in industry with the research and development people. Also research and development people, 90% of what they do is routine. It is the same sort of resistance. So I am convinced that 90% of their work is routine, it can be measured and it can be improved.

That prospective is difficult to sell. It really helps if you start with the manufacturing processes and the back offices; then if you have the results; it really works to convince people.

Companies who are beginning Six Sigma: what is the first thing they should consider?

At the companies that contact us because they want to start, my very first questions are always: why do you want to start Six Sigma; what is your vision? What do you want to accomplish with Six Sigma, why Six Sigma, and why now?

Because Six Sigma is not something you can do on the side; the scope and the amount of effort that it takes are quite intense. That means as senior management you really have to understand for yourself why you want to do this. If you don’t understand it yourself and you can’t explain it in a few lines, you can not explain it to your organization, and it is not going to work.

You send people home then if they don’t have an answer?

No, we’re businessmen. We help them create a vision! They have a sort of sense why they want to start, but you have to help them find out why and how they want to develop it.

Then you can take the next steps, which would probably be to appoint one or two programme managers, who are going to take care of the necessary practical arrangements. Ask the programme managers to select the first projects, for the first wave of Black Belts.

You have to think about human resource issues: do we draw Black Belts from the line organization or from the staff or from both; or do we want to hire external people as Black Belts? People who are made Black Belts, does that mean they get a new job or is it a combination of their regular job and a Black Belt job? So you have to make some decisions about that.

I try to refrain from doing too much planning, especially what I don’t plan before the first wave is a system for project tracking and too extensive change management models.

You have to get some experience in the company because I really adapt Six Sigma for the company for which I work, which means that I have to learn during the first wave about the company and the company needs to know about me.

After the first wave, it is a good time to design a system for project tracking and make modifications to the training curriculum, to think through a better system of project selection. But for me it doesn’t make sense to do that up front. Because then you are simply guessing what will work and what won’t. If you simply start and don’t make too much of a fuss about it, you have time to learn about the organisation and what works and what doesn’t work.

Each organisation has its own issues, its own problems and its own opportunities and it takes me about two or three months before I discover what they are.

The company culture.

Exactly! So you may have a company that has everybody defending the status quo and nobody defending change. Or you may have a company where everybody is working individually and there is no overall picture. That would be discovered after two or three months and you have to modify that.

I always explain to the participants of the first wave of training: they are the pioneers, and they have a hard job. A lot of facilities aren’t working yet. But my aim is to get it right for the second wave.

Also very important is if you start, make sure that you have some quick wins. Before you start rolling-out Six Sigma over the entire organization and before you make your change managements plans, you have to have some quick wins.
And that is why I start the big roll-out and the big communication plans after two or three months: first you need some proof that it really works.

That means in the pilot projects, I try the Failure Mode and Effect Analysis: that almost always gives some quick wins. That is very important. I find it a pitfall to do too much planning and communication upfront. Just start, make sure that you have some quick wins within two or three months that you have something to show and then you do the roll-out. That really works.

You talked about company culture, but there is also a big difference between the countries, the different cultures. How do you experience that?

Yes there is! I have delivered training in the US and China, it is really different.

For instance: typically in the Netherlands people are quite assertive. Often this comes down to being a bit on the anti-conformist side of the spectrum, being very individualistic. That is really different from China, where it is really hard for Black Belts, who are supposed to have an opinion of their own and take initiative.

In the Dutch situation, the sometimes excessive individualism needs to be managed, because it doesn’t work. There is something to gain if everybody follows the DMAIC steps, it gives a common language.

On the other hand, in China, it is difficult, they have to learn that it is nice to follow the guidelines, but they also need to be critical. They are trying to improve something so it is good to be critical. You won’t find something new by just following the rules and the guidelines.

You trained Black Belts in China successfully?

It is difficult to assess, because I only did the training and I wasn’t involved in project support. I was actually only hired to do the training of the Improve phase, the Design of Experiments part. It was quite hard for me to get interaction from them. They gave me a good evaluation but it is hard for me to tell!

Interview by Ingrid Vanlangenaeker